Re-optimizing an old stock portfolio means rewriting the metadata on files that are already live so they surface for searches they currently miss. Xpiks, StockSubmitter, Microstock+, PhotoTag.ai and Rastock AI all handle parts of this, but they solve different halves of the problem, and one agency rule limits how far any of them can take you.
This is one of the most common questions long-running contributors ask, and it is usually asked too late. A portfolio built over five or six years carries several generations of keywording habits, and the earliest files are typically described the worst while sitting on the strongest download history. Below is what each tool category actually does, where it stops, and how to sequence the work so you do not damage files that are currently earning.
What re-optimizing actually changes
Three things are editable on a live file at most agencies: the title, the description, and the keyword set. The image itself, its category assignment in some cases, and its accumulated download history are not things you can rewrite. That boundary defines the whole exercise. Re-optimization is a text problem applied to assets whose visual quality is already fixed.
The reason it works at all is that stock search runs largely on that text. A competent photograph described with eight generic keywords in 2019 is invisible against the same photograph described with thirty accurate ones today, and the gap widens as agencies lean harder on semantic matching.
Decide what is worth re-optimizing before you touch anything
Reworking ten thousand files uniformly is the wrong instinct. Most portfolios follow a steep distribution where a small share of files carries most of the income, and a large tail has never sold at all. Those two groups need opposite treatment.
- Files that sell steadily: leave them alone, or change one field at a time. Their metadata is already matching real buyer queries, and a wholesale rewrite can break a match you cannot see.
- Files with impressions but no downloads: highest-value target. The image is being found and rejected by buyers, which usually means the title promises something the frame does not deliver, or the file is competing in a category it does not belong in.
- Files with almost no impressions: second target. These are not being found at all, which is a keyword coverage problem rather than a relevance problem.
- Files that are genuinely weak photographs: skip them. No metadata rescues a picture buyers do not want, and time spent here is time not spent on the second group.
That triage depends on having per-file performance data in front of you, which is a separate capability from editing metadata and one that many tools simply do not have. Working out which keywords actually drive your downloads is the step that turns a re-optimization project from guesswork into targeting.
The tools, and what each is genuinely good at
These fall into four categories that get discussed as if they were one: desktop metadata editors, submission managers, managed pipelines, and tagging services. They are not interchangeable, and the right answer depends on where your files live and how many agencies you deliver to.
Xpiks
Best for contributors who keep everything locally and want direct control over embedded metadata. Xpiks is a desktop application that edits IPTC and XMP fields in the files themselves, with spell checking, keyword presets and export to agency formats. Pricing follows a paid desktop licence model rather than a subscription per file. The main limitation is that it operates on your local copies: if a file is already approved at an agency, Xpiks improves what you hold, not what the agency's search index holds, so you still need a separate route to push the change.
StockSubmitter
Best for high-volume contributors who submit to a long list of agencies and want one queue for all of it. Its strength is breadth of agency coverage and granular per-agency field handling, which is exactly what a large back catalogue needs. Pricing is subscription based on the vendor's own terms. The limitations are that it is Windows-centric, the interface rewards experience rather than being immediately obvious, and its metadata generation is not its main selling point. It is a distribution tool that also edits metadata, not the reverse.
Microstock Plus
Best for contributors who want submission management without a large upfront commitment. It covers uploading and submission across multiple microstock agencies and is frequently recommended in contributor forums for that reason. As with StockSubmitter, the re-optimization use case is secondary: it is built around getting new work out, so bulk editing an existing catalogue is possible but not the shape the product is designed around.
Wirestock
Best for contributors who would rather hand the whole distribution problem to someone else. It is a managed pipeline: you upload once, it handles keywording and submission to a set of agencies. The trade is structural rather than a flaw, and it is the one to be clear-eyed about. A managed pipeline operates on a revenue-share basis and takes a position between you and the agency, which is a poor fit for a back catalogue you already own outright at agencies you already have accounts with.
PhotoTag.ai
Best for contributors who only need better keywords and already have a delivery method they are happy with. It reads an image and returns titles, descriptions and keywords, with a credit-style pricing model. For a re-optimization project this is genuinely useful on the tail of a portfolio, where the existing metadata is thin. The limitation is scope: it produces text, it does not know each agency's field limits or banned terms, and it does not deliver anything anywhere. You still need the rest of the chain.
The agencies' own bulk editors
Best for small, surgical fixes, and free. Every major agency lets you edit titles and keywords on live files through its contributor dashboard, usually with some multi-select capability. This is the only route that changes what is actually in the agency's search index, so at some point every re-optimization workflow passes through it. The limitation is obvious once you try it at scale: the editors are slow, the selections cap out, and doing the same work five times across five dashboards is where most contributors abandon the project.
Rastock AI
Best for contributors who want the metadata layer and the multi-agency rule handling in one place. It generates titles, descriptions and keywords from the image, applies per-agency rules including banned and mandatory keyword lists before anything is written, delivers to 10+ agencies over FTP or SFTP, and keeps per-file status so you can see what landed where. It supports rollback, which matters more in re-optimization than in new uploads, and it exports IPTC, CSV and XML freely so nothing is locked in. There is no revenue share and no ownership claim over your files. The feature list covers the rule layer in detail. The honest limitation: for files already approved and live at an agency, it improves and delivers your metadata, but the agency's own editing rules still govern what can be changed on a published asset.
When to use a desktop editor, a managed pipeline, or a metadata layer
Use a desktop editor like Xpiks when your archive lives locally, you want embedded metadata corrected at the source, and you are comfortable driving the tool yourself. It is the right answer if the goal is to fix your master files once and never depend on a service.
Use a submission manager like StockSubmitter or Microstock Plus when the bottleneck is the number of agencies rather than the quality of the text. If your metadata is decent and the problem is getting corrected files back out to eight destinations, that is what these are built for.
Use a managed pipeline like Wirestock when you want to stop thinking about distribution entirely and accept a share of royalties as the price. For a back catalogue this is rarely the right trade, and it is worth understanding which tools take a revenue share and which do not before committing an existing portfolio to one.
Use a metadata layer with per-agency rules, such as Rastock AI, when the same archive has to satisfy several agencies with different limits and different banned terms, and when you need to be able to undo a batch that went wrong. Precision matters more than speed in re-optimization, because you are editing assets that already have earnings attached.
Use a tagging service like PhotoTag.ai when the only gap is the words themselves and the volume is modest. It is the cheapest way to test whether better keywords move anything before committing to a larger workflow change.
The constraint nobody mentions: you often cannot replace an approved file
Contributors often plan a re-optimization project assuming they can simply resubmit better versions of old files. Shutterstock's resubmission guidelines state the opposite: approved content cannot be overwritten with a new version. You would have to delete the original first, the replacement is reviewed against current standards with no guarantee of approval, and the download history and search ranking of the original do not transfer.
That single rule reshapes the whole project. Re-optimizing means editing the metadata on files that stay in place, not replacing the files. Any tool or plan that depends on deleting and reuploading your best-performing assets is trading a known earner for an unknown one, and no keyword improvement is worth that.
A safe re-optimization sequence
- Export your current per-file performance data from each agency dashboard so you have a before picture you can measure against later.
- Back up the existing metadata. Whatever tool you use, keep a copy of the current titles and keyword sets before overwriting anything.
- Pick one batch of one to two hundred files from the impressions-without-downloads group. Not the whole portfolio, and not your best sellers.
- Rewrite titles first, keywords second. Titles are where mismatch between promise and content usually lives, and they are the cheapest thing to change.
- Apply the changes at one agency only, and wait. Search index updates are not instant, and thirty to sixty days is a realistic window before the data means anything.
- Compare against your before picture. If the batch improved, roll the same approach out to the next group and the other agencies. If it did not, you have lost one batch rather than a portfolio.
Where most re-optimization projects fail
They fail on scope rather than on tooling. A contributor decides to fix everything, works through four hundred files by hand, loses momentum, and leaves the archive in a half-edited state that is harder to reason about than before. The counter is to treat it as a series of small measured batches, and to fix the underlying keyword habits rather than the individual files, so the same problem does not regenerate with next month's uploads.
The second failure is inconsistency across agencies. If you rewrite metadata for Adobe and leave Shutterstock on the old text, your own files start competing against different versions of themselves. Normalising keywords across the whole library first, then applying per-agency rules at delivery, avoids that split.
If you want to test this on a real batch before committing to a workflow, Rastock AI runs a 14-day free trial with no card required, and the pricing page sets out what each plan includes. Start with a couple of hundred files from the tail of your portfolio, where there is upside and very little to lose.
Frequently asked questions
Does re-optimizing old stock photos actually increase downloads?
It can, but only for files that were poorly described rather than poorly shot. The mechanism is straightforward: stock search runs largely on your title, description and keywords, so a file described with eight generic terms is invisible next to the same image described with thirty accurate ones. Files that are already selling well are the ones least likely to improve and most likely to be damaged by a rewrite.
Can I delete an old file and reupload a better version with new metadata?
You can, but it usually costs more than it gains. Shutterstock states that approved content cannot be overwritten and that the download history and search ranking of the original do not transfer to a new version, which is also reviewed against current standards with no guarantee of approval. Editing metadata on the file in place is the safer route in almost every case.
Which files should I re-optimize first?
Files with impressions but no downloads. Those are being found and passed over, which points to a mismatch between what the title promises and what the frame delivers, and that is fixable with text. Files with almost no impressions come second, since those need broader keyword coverage. Steady sellers should be left alone or changed one field at a time.
What is the difference between Xpiks and a tool like Rastock AI for this work?
Xpiks is a desktop editor that writes IPTC and XMP fields into your local files, which is ideal if you want your masters corrected and you handle delivery yourself. Rastock AI sits a layer up: it generates the metadata, applies each agency's rules and banned or mandatory keyword lists before writing, delivers over FTP or SFTP to 10+ agencies, and supports rollback if a batch goes wrong.
How long before I can tell whether the changes worked?
Allow thirty to sixty days. Agency search indexes do not update instantly, and download data on any single batch is noisy over shorter windows. Apply changes to one batch at one agency, record your before figures, and resist the urge to keep editing during the measurement period, otherwise you will not know which change produced the result.
Should I use the agency dashboard or a third-party tool?
Both, at different stages. The agency dashboard is the only thing that changes what sits in the agency search index, so every workflow ends there. A third-party tool is what makes the work survivable at volume: it generates and normalises the text, keeps it consistent across agencies, and gives you a backup to restore from if a batch performs worse than before.